Adoptive cell transfer (ACT) using genetically-engineered or tumor infiltrating T cells can have efficacy in patients with metastatic cancers, but the efficacy of ACT may be limited by the ability of T cells to expand and persist following transfer. T cells harvested from blood or tumor for ACT are often comprised of memory and late effector subsets. Despite phenotypic and functional heterogeneity, all CD8+ memory T cell (TMem) subsets express Fas, a tumor necrosis factor receptor superfamily member conventionally known as a death receptor. Because Fas can mediate non-death signaling in other cell types, we hypothesized that a cognate Fas-FasL interaction within the tumor microenvironment might limit both T cell persistence and anti-tumor efficacy. We discovered that FASLG, the gene encoding the apoptosis-inducing ligand FasL, is overexpressed within the majority of human tumor microenvironments. We tested whether blockade of Fas-induced memory T cell differentiation could improve therapeutic outcomes in multiple adoptive immunotherapy models. Augmentation of Fas signaling in TMem using an oligomerized form of FasL enhanced cellular differentiation and induced apoptosis. Conversely, antibody blockade (anti-FasL) of Fas signaling in TMem culture in vitro retarded phenotypic and functional differentiation, resulting in fewer exhausted cells and higher T cell expansion in vivo. Cell-extrinsic Fas blockade using anti-FasL provided superior anti-tumor efficacy in multiple syngeneic murine tumor models but normalizing for differences in T cell differentiation erased these improvements over IgG controls. Genetic engineering of Fas variants impaired in the ability to bind FADD functioned as dominant negative receptors (DNRs) and allowed us to block Fas signaling in a cell-intrinsic manner. This rescued Fas-competent mouse and human T cells from FasL-induced apoptosis. Fas DNR-engineered T cells exhibited enhanced persistence within tumors following ACT, resulting in superior cancer regression and overall survival. Despite enhanced longevity, Fas DNR-engineered T cells did not undergo aberrant clonal expansion, demonstrating the safety of this approach. Thus, cell-intrinsic ‘insulation’ of T cells from the negative influence of FasL is a potentially universal strategy to enhance ACT efficacy across a broad range of human malignancies.